A Metaphysics for Wandering Coyote

Posted on Mon 13 September 2010 in Rumination

stock photo of coyote

On the Origins of Objects

For lack of opportunity I do not get to read and learn all that I would want to, or might need to. And so, in all but a few private conversations, I have hesitated to voice some of my most visceral metaphysical intuitions. And yet, for one reason or another, I have been drawn down the path toward deeper and deeper engagement with metaphysical problems, no matter how I try to pull back from them.

In many ways then it is a relief and a satisfaction to read Brian Cantwell Smith so eloquently give voice to very similar intuitions in his book On the Origin of Objects published in 1996 by MIT Press. It is beautifully written and cleverly illustrated book and has Daniel Dennett's praise, calling it "as deep and original as any book by a philosopher (or computer scientist) on the topics of representation, ontology, and meaning." Dennett's praise is reassuring, in the face of the Smith's ambition to do no less than to create a new successor metaphysics for the 21st century.

Whether or not Smith succeeds in this project is an open question that I am not qualified to answer. Having said that, and considering that Smith's metaphysics has been taken seriously by some respected thinkers, I am quite willing to use his work as a kind of inspirational platform from which to launch a particular sort of investigation of environmental and semantic information... and, more lamely, as an excuse not to have to tread such deep metaphysical waters trolling for definitive answers to impossible questions. Fortunately, an entirely perfect metaphysics is not a pre-requisite to what I hope will be a useful and illuminating exercise.

I will briefly describe Smith's views, though it will not do justice to them. A much more thorough overview can be found by Josefa Toribo's very helpful review(pdf) of Smith's book (reference at bottom of page). Those interested in reading Smith's book, may do well to read Toribo's article first.

My goal is to use Smith's approach to metaphysics to motivate a way of framing questions about information, and in particular how the objects of situation theory's ontology might be modeled*. I am not certain where such an investigation will end up. I'm starting here, but ending where I do not know exactly, not yet.

Smith's Thesis

Individuality vs. Particularity

Smith distinguishes between particularity and individuality. An individual is an object, a thing that is taken to be separate from the world, having various properties, and supporting re-identification over time and space. Individuality is that which allows us to count. Particularity is the property of being a situated "patch" of "metaphysical flux" (like a situated patch in space time). Particularity enables specificity. One can have particularity without individuality (e.g. when the material world is not being individuated) , individuality without particularity (e.g. types, laws, kinds, properties), and bare particulars which have both particularity and individuality**. It is to the notion of individuality that the notion of object is most closely connected, paradigmatically in the form of things that have both particularity and individuality (like my old work boots).

A physical metaphor

Smith divides physics into three realms: The first realm is the material realm of the particular, the underlying metaphysical flux that makes up the world. The second realm is the everyday realm of practicing physicists. The third realm is the universal realm of laws, kinds, and types, filled with abstract individuals lacking particularity. The world of the first realm flows from one state of deictic metaphysical flux to another, guided by localized applications of universal laws found in the third. Unlike logic, physical laws do not explicitly quantify over individual objects. Instead, physical laws are given entirely in terms of the realm of universals, which govern the realm of the particular. Since "nothing non-particular is itself governed by physical law" (p. 156), physics is not equipped to transform the world's continuous deictic fields of particularity into ordinary sorts of objects like stars, tables, rocks, and lobsters, possessing individuality in addition to particularity:

Ontologically, in particular, it is not clear that physics actually does presuppose the notion of individuality. That is, it not clear that the pure physical (first) realm is assumed by physics to be populated with individual objects at all. This is easiest to see on a field theory interpretation...The picture of (first-realm) world to which this conception is committed is one of infinitely extensive continuous fields of mass, charge, etc., ebbing and flowing in dynamic ways as governed by the relevant classical laws. Given the results of the previous section we can say that the world of classical physics, at least on this interpretation, is one of infinitely extensive continuous, particular, deictic fields. p. 176-7

And a little further, he writes:

In sum, being an individual object is also not, in and of itself, an effective or even salient physical property. No physical attribute holds of an individual, for example, except in virtue of its physical composition...For there are no physical objects. (p.177-8)

This is not to say that the things that we take to be objects do not have a material basis about which physics can inform us. Rather it is a comment on the role physics has to play in bringing out of a continuous metaphysical flux discrete objects. In a world where what materially exists is an objectless unity of "infinitely extensive continuous fields" of particularity, "physicists will have to look to a theory of intentionality for the notion of an individual, not the other way around," (p. 180).

Nonetheless, physicists frequently refer to objects in their work. What is going on? Smith argues that this gap between how physical theory registers the world and how physical theory is applied in practice is a matter of epistemic, computational, and practical necessity (p. 179-181). Objects are intentional achievements in a way that the underlying material world is not.

From this Smith derives a stringent methodological moral:

Criterion of Ultimate Concreteness

No naturalistically palatable theory of intentionality—of mind, computation, semantics, ontology, objectivity—can presume the identity or existence of any individual object whatsoever. (p.184)

In this way, Smith seeks to

make room ... to recognize that objects are inexorably cultural, biological, political, psychological, social, evolutionary, historical, economic, and so on and so forth—i.e., are everything that social constructionism and critical studies have shown them to be—without, in the first(aboriginal) instance, having to pay the price of these ineliminable but nevertheless very expensive notions.(p. 188).

For those attracted to constructivist metaphysics, but are uncomfortable with some of the costs it exacts, this middle-road, has an obvious attraction.

Individuals in and individual-less world

A monk, taking a bamboo stick, said to the people, "If you call this a stick, you fall into the trap of words, but if you do not call it a stick, you contradict facts. So what do you call it?" At that time a monk in the assembly came forth. He snatched the stick, broke it in two, and threw the pieces across the room.(source)

The question then is how to build up individuals out of an individual-less world. His answer, more-or-less, is participatory processes of registration within a loosely-coupled world permitting both linkage and separation of effect between regions.

smooth surface

Registration is defined as something like the act of taking something to be something. It is it "parse, make sense of as, find there to be, structure, take as being a certain way—even carve the world into..." (p. 191). It is what happens when one region of metaphysical flux, tentatively designated as the s-region, retains orientation toward another region of flux, tentatively called the o-region, even when those two regions are no longer effectively connected. It is distinguished from perception by the fact that perception ceases as soon as the s-region is no longer effectively coupled from the o-region. If my daughter goes into the other room, I can no longer perceive her, but I still register her presence despite the separation.

Registration is intentional, it is directed, and the responsibility for maintaining orientation is asymmetrical. The s-region must stabilize the o-region (and itself) as objects. The world is not so tightly coupled that change in one thing will require a change in all other things. Rather, effect is dissipative, just as a gravitational field weakens with distance from a mass. This loose coupling of the world is what Smith calls flex and slop (p. 191-212). This flex and slop is what permits the decoupling of registrar from registered, and which according to Smith, is the basis for individuation and, ultimately, intentionality. In order to retain orientation toward something from which one has been effectively decoupled requires that the target of that orientation be stabilized into something separate and re-identifiable, an abstraction of shifting particularities across time and space. The usefulness of this is pretty clear, as when a prey animal tracks a predator behind some bushes, or predator its prey momentarily hidden in the tall grass, or when a truck driver keeps the motorist that moved into his blind spot in mind. But this abstraction, separation, and flex and slop also introduce the possibility of error and uncertainty.

It is this quasi-realist quasi-constructionist that I find so appealing: that of an object-less, continuous, material world exhibiting various properties, which we individuate as things like mass and energy, upon which an epistemic agent is free*** to register or carve out any arbitrary region(s) of that world to treat as if they were objects, on the fly. Nonetheless, this process is not one simply of agents perceiving the world in terms of objects:

Because the ineliminable relations that connect the web of practice to what is registered are in part connected—implying that a kind of ineliminable and unregistered "effective connective tissue" holds up every registerable thing, every act of registration—it follows that registration is not simply a process of selecting or filtering the world. On the contrary, it may, and usually will, also involve shaping the object in such a way that it can be seen as world. Participatory processes are as much violent as they are descriptive: we adjust the objects so that our theories are true of them, as well as adjusting our theories so that they are true of the objects. (p. 299-300)

We must be careful to register that not all ways of carving up the world into objects is equally useful for every purpose, or have identical properties. The material world, the metaphysical flux, is not itself uniform, and patterns within that flux can be of varying complexity, with some individuations more coherent than others. Moreover, not all individuations have the potential of contributing equally to the business of livingS, some more useful to agents having to solve the practical problems of encountering the brute reality of the physical world. Natural selection cares not for metaphysics, but responds to the way the way the world extrudes itself upon itself, in the way sticks that are not sticks get broken.

Taking the World Seriously

This picture relates in a way to a fundamental concern of mine that we take seriously the notion that life, intelligence, intentionality, and selection are embodied processes that emerge naturally from the world itself- that instead of modeling an organism as something that has a pre-defined set of properties and methods, i.e. as an individual object, that we find a way to model organisms not as individuals, but as material realities that emerge from the background structure and operation of its world. For this reason, I have found the constructions within cellular automata like Conway's game of life so intriguing. More on that later.

I have given only a brief introduction to these ideas, and I am sure that I have ran into more than a few potholes, traps, and other philosophical quagmires along the way. But the basic picture appeals to me, and I've got my work boots on to trek through the mud if have the mind to.

Works Cited

Smith, Brian C. On the Origin of Objects, MIT Press, London. 1996

Toribo, Joseph "Extruding Intentionality from the Metaphysical Flux" in Journal of Experimental and Theoretical AI, 11, 1999: 501-518. http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.100.5447&rep=rep1&type=pdf

* Too many good research programs have broken down for no reason but the accidents of history, having fallen on soil not yet fertile, or suffered misfortune before full bloom. I am not convinced that progress is the inevitable consequence of scholarship, especially if we let the work of those of generations before us fade from memory and fall from bookshelves into waste-baskets.

** This usage of bare particular is somewhat non-standard, and the distinction being made is, according to Smith, not usually made. In metaphysics there are two major camps regarding the nature of objects: the substratum theorists (or substance-attribute theorists), and bundle theorists. The substratum theory holds that an an object has both thin (or bare) particulars and properties (or universals) and that these are glued together under the relation of instantiation. Hence, if two objects have exactly the same properties, they need not be identical, since they needn't have identical bare particulars. The nature of such particulars, it is said by some, is somewhat mysterious. On the other hand, bundle theorists hold that objects are nothing more than bundles of properties under the relation of compresence. Stripped of all properties, nothing remains. Bundle theory seems to imply that if an object is identical in all its properties to another, then they must be the same object. This is seen by some as problematic. Similar concerns about identity have arisen in modeling situations in situation theory, particularly when situations are modeled mereologically as sets of infons or types. That discussion will have to wait for another occasion.

*** Free that is from metaphysical constraints. Practical, cognitive, and physical constraints are to be expected.